The rising adoption of electric vehicles (EVs) utilizing lithium-ion batteries necessitates a robust understanding of state-of-health (SOH) estimation. The existing literature highlights various SOH estimation models, but a comprehensive comparative analysis is lacking. This paper addresses this gap by conducting an exhaustive review of diverse SOH estimation approaches for EV battery applications, including the direct measurement method, physical-based and data-driven approaches. Results highlight that data-driven methods, particularly those utilizing machine learning techniques, offer superior accuracy and adaptability but often require extensive datasets. In contrast, physical-based approaches provide interpretable insights but are computationally intensive, and direct measurement methods, though simple, lack generalizability. In addition, this paper also systematically reviews the indicators of battery SOH, influential factors affecting battery SOH, and various datasets used for SOH modeling. Future research should focus on integrating multiple modeling methodologies to leverage their combined strengths, enhancing the collection of comprehensive battery lifecycle datasets to support robust model development, and extending the scope of SOH estimation beyond individual cells to encompass entire battery packs.

To improve the energy efficiency of hybrid commercial vehicles under varying road gradients, this study develops a road-gradient-aware hierarchical energy management framework. An interactive multiple-model Kalman filter (IMM-KF) is employed to estimate the current road gradient by fusing information from vehicle dynamics and onboard measurements. At the vehicle level, an MPC-based strategy uses future route-gradient information extracted from a pre-reconstructed road profile, rather than predicted by the IMM-KF, to plan an energy-efficient speed trajectory and reduce vehicle energy demand. At the powertrain level, a particle swarm optimization (PSO)-calibrated adaptive equivalent consumption minimization strategy (PA-ECMS) allocates the resulting power demand between the engine and motor, with the SOC-feedback PI parameters calibrated offline. The framework is evaluated in MATLAB R2024a/Simulink using standard driving-cycle and real-road data-based simulation scenarios. The IMM-KF achieves an RMSE of 0.04461° under constant-gradient conditions and 0.47587° under variable-gradient conditions. The upper-layer MPC reduces wheel-end energy consumption by 10.98% compared with PID-based speed control, while PA-ECMS reduces the comprehensive operating cost by 15.3% relative to A-ECMS under the prescribed CHTC-LT cycle. In the integrated real-road simulation, MPC + PA-ECMS reduces fuel consumption and comprehensive operating cost by approximately 4.0% and 4.7%, respectively, compared with PID + A-ECMS. These results demonstrate the effectiveness of coordinating road-gradient estimation, economic speed planning, and hybrid power allocation within the proposed hierarchical framework.

Vulnerable road user (VRU) detection in complex traffic scenes is challenged by the degradation of weak local details during deep feature propagation, background-dominant responses introduced by global-context aggregation, and insufficient differentiation of localization quality during regression and positive-sample assignment. To address these issues, this paper proposes GPL-YOLO. First, C3k2-GSB preserves fine-grained texture and edge information through an identity branch while enhancing spatial and global–local representations through cascaded GSB units. Second, Polarized Linear Attention (PLA) combines efficient global interaction with polarization mapping and gated modulation to suppress background interference during multi-scale feature fusion. Third, Focal-Wise IoU is coupled with improved Task-Aligned Label Assignment (TAL) to introduce localization-quality-aware optimization into both bounding-box regression and positive-sample selection. On KITTI_VRU, GPL-YOLO achieves 90.5% Precision, 78.6% Recall, 87.4% mAP50, and 55.8% mAP50–95, with 3.093 M parameters and 7.8 GFLOPs. Additional evaluation on BDD100K_VRU shows consistent improvements over YOLOv11n under the shared Pedestrian/Cyclist label space. These results demonstrate the effectiveness of the proposed complementary improvements in feature preservation, contextual modeling, and localization-quality-aware optimization.